1. Field of the Invention
The present invention relates to an image processing apparatus, an image processing method, and a program, which are capable of processing an image of a measurement object to measure the size of the measurement object.
2. Description of Related Art
In the past, measurement of a blade or the like in a jet engine has been performed using an observation tool such as an endoscope. Japanese Unexamined Patent Application, First Publication No. H3-102202A discloses a technique applied to measurement of the blade or the like. In the technique disclosed in Japanese Unexamined Patent Application, First Publication No. H3-102202A, a subject image obtained by imaging a subject and a CG image generated by Computer Graphics are displayed on a monitor. Then, the CG image is changed based on a user's instruction so that a state of an object in the CG image matches a state of a subject in the subject image. Thus, according to the technique disclosed in Japanese Unexamined Patent Application, First Publication No. H3-102202A, it is possible to measure a desired portion of the CG image.
However, in the technique disclosed in Japanese Unexamined Patent Application, First Publication No. H3-102202A, when measurement is performed on each subject image while switching a plurality of subject images which differ in appearance of a subject, an operation of matching the state of the object in the CG image with the state of the subject in the subject image needs to be performed each time a subject image is switched.
Meanwhile, a technique of automatically matching a 2-dimensional (2D) image obtained by imaging a subject with a 3-dimensional (3D) object based on data obtained using a computer-aided design (CAD) is disclosed in U.S. Patent Application, Publication No. 2009/0096790. In the technique disclosed in U.S. Patent Application, Publication No. 2009/0096790, first, when a 3D object is observed from a plurality of virtual points of view, projection diagrams of the 3D object observed at the respective points of view are generated. When the projection diagrams are generated, the plurality of projection diagrams are stored together as a 2D model. Then, matching between the 2D image and the generated 2D model is performed. Then, a projection diagram which is highest in similarity with the 2D model is searched for, and a relative position of a subject shown in the 2D image and a camera that has imaged the 2D image is calculated based on the position of a point of view when the projection diagram is generated.